Wednesday, February 1, 2012

Have journals become redundant and counter productive?

"Is this [publication in high profile journals], then, an efficient way to marshal evidence in support of grant proposals, appointments, promotions or fellowships? Of course not - it is madness. The time is overdue to abolish journals and reorganise the way we do business"

David Mermin, Publishing in Computopia, Physics Today, May 1991

[This was written a few months before the arXiv started. See also the followup column, What's wrong in Computopia? in April 1992 which was after the arXiv started]

Given the way the world is today, rather than one hundred years ago, if we wanted to design the most effective way to promote good science would we come up with the current system of scientific journals?

Keep in mind the following:

Journals consume vast amounts of resources including
- as much as 60 per cent of the total budget of many university libraries
  [see Academic publishers make Murdoch look like a socialist].
- the time of referees and editorial board members

The "best" journals such as Science and Nature
-skew fields and papers towards `sexy' and grandiose claims
-fail to stop the publication of fraudulent science such as that of Henrik Schon or of papers which fit 20 plus data points to a curve with 17 parameters

The arXiv has now been going for twenty years [see this commentary by the founder Paul Ginsparg in Nature last year]. Functionally, physicists almost exclusively use it to obtain and distribute scientific information independent of the peer review, ranking, and impact factors of journals. It is hard to make a case that this has led to a decline in scientific standards in the physics community. [Why is there no arXiv in chemistry? has no clear answer].

When using the arXiv, I anticipate people make their own judgements of the importance, validity, and significance of papers largely based on the actual scientific content of the paper. I suspect the reputation of the authors also comes into play at times. But I would contend that reputations are still largely based on actual scientific track record rather than on more debatable criteria such as numbers of Nature papers.

So why do journals still exist? Perhaps the two main reasons are:
- the vested interests from publishers
- the institutional inertia within funding agencies and employers who still consider them essential for ranking applicants for funding, jobs, or promotion. One might argue that this inertia is driven by intellectual laziness (i.e. bean counting).

In an ideal world, abolition of journals would free up significant resources that could be used instead to
-employ more people to actually do the science, including to write good review articles that meaningfully assess, filter, and critique the vast seas of scientific information
-provide time for people to make assessments of applicants based on their actual scientific achievements rather than their performance on metrics.

But, how do we get there? Will we ever?


  1. I personally hate looking things up in certain journals (meaning certain publishers) and really only read stuff that's on the arxiv or papers worth the trip to the library (a good example, is that i recently looked up some of rashba's old work and that quality is work the walk).

    i also think ive seen flat out wrong papers in PNAS and nature/science are usually vague and overblown. people frequently lament how their good work was rejected from X but published in Y (hey, PRL see this?)

    when i write citations in my papers i cite the papers that i found useful to me. then usually my coauthors make me site the 'important' people. even if they suck.

  2. I wonder if a world with arXiv alone would work.

    It's true, I check the arXiv daily, but the PRL website once every week or two, prb even less, and Nature/Science very occasionally (never ever Nature : Physics, not sure exactly why except that the trip always seemed pointless in the end). That being said I'm keen to publish in any of those!

    But without the inevitable review process, I wonder if arXiv submissions would become sloppy and unpolished, mistake-ridden, and more often wrong.

    That being said, one's reputation is at stake with arXiv submissions, and they're more risky having not been reviewed, and so perhaps it self regulates quite well.

    What I really love about the arXiv is that journal prejudice, on new submissions, is completely absent (except hints: 4+ pages? They're going for PRL!), and so as you say, scientific merit and relevance to my current work alone determine my interest.

  3. In answer to the "arXiv in chemistry" question, it would be interesting to get some data on the frequency of patent filings related to published work. You can't publish before patenting, so for many experimental chemistry labs there will be a delay anyway between the time the patent is written up and the time the work is published. Of course, there is plenty of work in physics with direct application that could be patented, so it is not clear that is really is the key to understanding the cultural difference.

    My PhD advisor remarked to me once that chemistry has traditionally been the most money-driven of the natural sciences. This goes back to the alchemists, really, whose goal was literally to make money...